4 ms·
That may be true, but they are presumably already maintaining a bunch of #ifdef's for the Windows-only features.
by pyre 13y ago
That may be true, but they are presumably already maintaining a bunch of #ifdef's for the Windows-only features.
- eridius 13y agoGiven the relative popularity of Windows vs Linux, it probably is worth the effort to maintain Windows-only features.
- 67726e 13y agoTrue, but Windows has the much larger market share so it becomes worth it.
- kbenson 13y agoBeing able to assume the #ifdef for features go one direction can have some benefit in reasoning about the code it relates to. In other words, there may be benefits to keeping the windows features a superset of the linux features, most likely in preventing stupid programmer errors.
- chris_wot 13y ago> That may be true We'll never know. That's a binary blob on your machine.
- pyre 13y agoThis is true, but I fail to see the relevance here. This post just comes across as a "any chance to mention Free Software" type post. This thread is talking about if this decision by Nvidia makes sense. Trying to inject "well, we could all just ignore Nvidia if you were using Free Software," doesn't really contribute much to that conversation. > That's a binary blob on your machine. On my machine? It's less likely than you might think.
- subway 13y agoIt goes (or at least at one time went) well beyond that. Last I heard, nvidia.o was shared across Linux, FreeBSD, Solaris, and Windows.
- ErsatzVerkehr 13y agoDon't these systems use different object-file formats?
- subway 13y agoAt the very least I can confirm the object is identical (same md5sum) across Linux, FreeBSD, and Solaris. Past sources have indicated to me Windows shares his object.